This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: New parameters to control stringop expansion libcall strategy
- From: Xinliang David Li <davidxl at google dot com>
- To: Jan Hubicka <hubicka at ucw dot cz>
- Cc: "Joseph S. Myers" <joseph at codesourcery dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, Teresa Johnson <tejohnson at google dot com>
- Date: Fri, 9 Aug 2013 12:41:23 -0700
- Subject: Re: New parameters to control stringop expansion libcall strategy
- References: <CANtU07_Oi2xTDX_K71MEpgkGBFXAVXtLGmdTqPNX3Yf=C7tHqA at mail dot gmail dot com> <20130806094210 dot GC20805 at kam dot mff dot cuni dot cz> <CAAkRFZL96-tNbQxtB16cUaPN7OHAFNa7rTF1pfpKP-+GMWK1_w at mail dot gmail dot com> <Pine dot LNX dot 4 dot 64 dot 1308080018460 dot 26185 at digraph dot polyomino dot org dot uk> <CAAkRFZKDiQW-gL0PJY-A1fPtDa-cAXi6SS38VFFaqWVgqDgmKQ at mail dot gmail dot com> <Pine dot LNX dot 4 dot 64 dot 1308080103210 dot 26185 at digraph dot polyomino dot org dot uk> <CAAkRFZJBjdq1gDw39U8MVyS8zsNMgO45XaQxFYiT3rWNX_HBOQ at mail dot gmail dot com> <Pine dot LNX dot 4 dot 64 dot 1308081517260 dot 7708 at digraph dot polyomino dot org dot uk> <CAAkRFZLiMtHuo=UCiFuAc0c+WT3BDUKmkP35+K9s1YuwFV+pVw at mail dot gmail dot com> <CAAkRFZKyJL47hMcggHyFUHJ-basgGj5q+oJi_XimdS9AWOSA9g at mail dot gmail dot com> <20130809183340 dot GD6579 at kam dot mff dot cuni dot cz>
On Fri, Aug 9, 2013 at 11:33 AM, Jan Hubicka <hubicka@ucw.cz> wrote:
>> Is this version ok for trunk?
>
> It looks resonable, but I still do not like much the removal of const for tables.
> Doing so will push them all into David Malcom's per-thread global universe.
>
> Currently the algorithm is selected based on cost->memset/cost->memcpy.
> Instead of removing the const of all the CPU tables, I would preffer
> introducing two readwrite global variables memset_algs/memcpy_algs and feed
> them with proper table at a time we set up ix86_tune_features.
>
I can do that in this patch. In the future, when we need to do tunings
for those constants, we can revisit it.
thanks,
David
> This has chance to do the right thing with optimize attribute specifying algorithms
> and with the longer term threading plan.
>
> Honza
>>
>> thanks,
>>
>> David
>>
>> On Thu, Aug 8, 2013 at 9:31 AM, Xinliang David Li <davidxl@google.com> wrote:
>> > Updated.
>> >
>> > thanks,
>> >
>> > David
>> >
>> > On Thu, Aug 8, 2013 at 8:18 AM, Joseph S. Myers <joseph@codesourcery.com> wrote:
>> >> On Wed, 7 Aug 2013, Xinliang David Li wrote:
>> >>
>> >>> Updated patch attached (fixed header, buffer overflow, and warning -->
>> >>> error problems).
>> >>
>> >> You still have diagnostics starting with a capital letter, contrary to the
>> >> GNU Coding Standards.
>> >>
>> >> --
>> >> Joseph S. Myers
>> >> joseph@codesourcery.com